Celtic family tree

This is the most heavily annotated diagram on the site, and it is the only one that has to begin with a warning: any Celtic family tree that looks tidy is fiction. The material comes from two separate literatures, neither of which was trying to produce a genealogy of gods, and both of which were written by Christians recording stories about beings their own frame denied were divine.

Two literatures, not one

The diagram is drawn in two zones, because Irish and Welsh figures do not connect.

The Irish side is dominated by the Tuatha Dé Danann, the people of the goddess Danu, whose Book of Invasions genealogy makes them one of a sequence of peoples who settled Ireland. The Dagda, the Morrígan, Lugh, Nuada, Brigid, Dian Cécht, Goibniu and Ogma belong here. Relationships between them are given in the manuscripts, frequently more than once, and frequently differently.

The Welsh side has the children of Dôn — Gwydion, Arianrhod, Gilfaethwy, Amaethon — and the separate family of Llŷr, with Brân, Branwen and Manawydan. There are tempting resemblances across the Irish Sea: Dôn against Danu, Llŷr against Lir, Manawydan against Manannán, Lugh against Lleu. Some of these are linguistically defensible and some are not, and none of them justify drawing a line between the two zones. The diagram keeps them apart and notes the parallels rather than merging them.

Danu, who may not exist

The name of the Tuatha Dé Danann is usually translated "people of the goddess Danu", and Danu herself appears almost nowhere. Her name is largely a back-formation from the genitive in the group's title. Some scholars prefer to read the phrase as the people of the god or of skill. Putting a matriarch at the top of the Irish tree, as many popular charts do, gives the whole structure a foundation the sources barely support, and this diagram marks the node as reconstructed.

Where the contradictions are

Irish sources contradict each other about parentage as a matter of routine, and the annotations carry the detail. Some of the most consequential:

• Lugh is the son of Cian of the Tuatha Dé and Ethniu, daughter of the Fomorian king Balor — making the hero who kills Balor his own grandson's victim, which is the point of the story rather than a confusion.
• Brigid is a daughter of the Dagda in the Mythological Cycle, and is also described as three sisters of the same name with three different specialisms. Whether she is one goddess, three, or a triplication is unresolved, and her overlap with Saint Brigid of Kildare is a separate question again.
• The Morrígan is variously a single figure and a group of three — with Badb, Macha and Nemain appearing in different combinations in different texts, sometimes with the Morrígan as one of them and sometimes as their collective name.
• The Dagda's children vary by source: Aengus, Brigid, Bodb Derg and Cermait are commonly given, and the lists do not match.
• Aengus' conception involves the Dagda stopping the sun for nine months so the pregnancy can pass in a single day, which is a story about concealment and produces a genealogy nobody can date.

Fomorians are not a separate species

The Fomorians are the Tuatha Dé's enemies and are also, repeatedly, their relatives. Lugh is half Fomorian; Bres, who rules the Tuatha Dé badly, has a Fomorian father and a Tuatha Dé mother; marriages across the line are how the two second battles are set up. Like the Norse giants, they are an opposing group rather than an opposing kind, and the diagram draws the crossings rather than treating them as anomalies.

The euhemerist frame

Everything on the Irish side is filtered through the manuscripts' insistence that these are people, not gods: a race who came to Ireland, were defeated by the Milesians, and withdrew into the sídhe, the hollow hills. That framing gives the material a pseudo-historical chronology, complete with dates and descent from Noah, which sits on top of the older material like a coat of paint.

The diagram follows the relationships the manuscripts state while marking the pseudo-historical layer as a separate register, because a reader needs to know which lines come from a story and which come from a monk fitting the story into biblical time.

Reading the diagram

Solid lines are relationships given consistently in the medieval sources. Broken lines carry a note naming the manuscript or cycle that disagrees, and there are more of these here than on any other tree. Reconstructed nodes — Danu above all — are marked as reconstructions. Where a figure appears in both a divine and a heroic register, as several do, both placements are shown.

Heroes of the Ulster and Fenian cycles are drawn where the sources give them divine parentage, which is how Cú Chulainn attaches to Lugh, and left unattached where they do not.

For why the evidence takes this shape — no contemporary writing, monastic transmission, and a century of romantic invention layered on top — see the Celtic tradition hub.